The NII will on one hand build on the Internet and the WWW. On the other hand, much of the infrastructure will be motivated by some early business opportunities, which roughly can be thought as digital versions of today's cable TV. As shown in Figure 3, several industry sectors are teaming and competing for this market, which also includes interactive shopping and the digital equivalent of the current video rental business.
Figure 3: The convergence of industries and capital investment needed
to deliver digital TV service
In Figure 4, we try to quantify this vision assuming that some 100 million clients would be connected to the digital media information system. The year in which this becomes reality depends a lot on both technology and government---some regulations could stifle competition and slow digital deployment; others would speed the process through encouragement of provision of digital services to remote areas that might be uneconomic in a purely market driven process. The NII will link hybrid delivery technologies (including ISDN, cable, ADSL, wireless, mobile systems) for the off-ramps to ATM and satellite, which can be used for both off-ramps and the long distance trunks. Further, we expect large-scale Massively Parallel Machines (MPPs) to be part of a network of video servers, and more generally, WebServers that supply information to these clients. We term this scenario InfoVISiON for Information, Video, Imagery, and Simulation on Demand. This includes the storage, query, and dissemination of this wide range of multimedia data. There is surely at least 100,000 hours of interesting video material in the archives of Hollywood studios, CNN, Routers, and network TV. If compressed in MPEG format, this corresponds to some 100 Terabytes of needed storage capacity. MPEG2 or other format, such as motion JPEG, would require much more storage and are probably necessary to support editing and other video production applications. Note that the current WWW has only a few percent of its storage devoted to video---the future NII will be dominated by video data. Each NII offramp will, as shown in Figure 5, connect homes (offices, school desks) at the rate of 1--20 Megabits/sec to the set of NII InfoVISiON servers. This rate covers the range from compressed VHS to HDTV picture quality. Note that this performance is 100--1,000 times greater than today's conventional 14.4 Kbaud modem on a twisted pair (plain old telephone service POTS) connection.
Figure 4: An estimate of the communication bandwidth and compute
capability contained in the NII and supercomputer industries
Figure 5: The basic InfoVISiON scenario as seen by a home in the year
2000 with an intelligent settop box interfacing the digital home to a
hierarchical network of InfoVISiON servers
Returning to Figure 4, we estimate comparable investment in clients and servers and find an InfoVISiON or WebServer market that is at least an order of magnitude larger than that for supercomputers. This illustrates that it is the NII, and not large-scale number crunching, which is the best opportunity for parallel processing. Notice also that the compute power contained in this future NII is some 10--100 PetaFLOPS---far larger than the compute capability of an individual TeraFLOPS supercomputer. This motivates our interest in WebWork, described in Section 3, and aimed at extending Web to compute servers, and harnessing the power of the World Wide Metacomputer.
One could view this InfoVISiON scenario as the most amazing
client-server application with 
clients and 
large
servers. However, Vice President Gore has articulated NII democracy
with everybody able to host information. Further new technology, just
as Java and WebTools (Section 3), suggests instead a
heterogeneous server-server architecture with 
distributed
nodes. This can be viewed as a fascinating parallel computer with
many more nodes than traditional tightly coupled systems (by a factor
of 
compared to typical large 1,000 node MPPs). It appears that
although powerful, the communication backbone will not allow every
``client'' long distance simultaneous access to every other client or
server. Rather, we must enforce the guiding principle of all computer
architectures---namely, data locality. This is illustrated in
Figure 6. When Jurassic Park VI is released on the
Hollywood Server, one will not have everybody accessing it there at
an average distance of some 1,500 miles. Rather, this ``hot'' movie
will be cascaded down through the hierarchy of servers so that any
individual will find it on a server a few miles away. This strategy
reduces the needed fiber for the NII trunks by a factor of about 100.
Figure 6: A typical hierarchical server network depicted for a master
system in Hollywood cascading down with a fragment of node systems
shown for central New York
Figure 7 illustrates a related point. We consider a WebServer as a system implementing the software model of Figures 1 and 2. This could be a server with full input and output access to the World Wide Information, or it could be aimed at a smaller organization with firewalls, as necessary, to enforce security. We see that the mix of distributed and conventional (PC, Mainframe) software of Tables 1, 2, and 3 makes WebServers the most attractive technology to build a business enterprise support system.
Figure 7: A Web server can be used at any level of organization from an
individual home to the entire universe
Table 3: Some emerging high-level WebWindows integration concepts